Skip to content

Add incident and operations extractor for PagerDuty and OpsGenie #506

@ravisuhag

Description

@ravisuhag

Context

When an AI agent is investigating a data quality issue, knowing that the upstream service had an incident last night is essential context. Incident history connects operational events to data assets.

Scope

New extractors for incident management platforms:

PagerDuty

  • Extract incidents: title, status, severity, service, timeline
  • Extract services and their escalation policies
  • Extract on-call schedules
  • Emit relationships linking incidents to affected services and data assets

OpsGenie

  • Extract alerts and incidents: title, status, priority, tags
  • Extract teams and on-call schedules
  • Emit relationships where inferrable from tags or service mappings

Design Considerations

  • Time-windowed extraction — only pull recent incidents (configurable lookback)
  • Relationship inference: match incident service names to known assets
  • Incident metadata should include resolution status and duration

Why

Operational context is critical for AI reasoning about data trust. "This table's freshness dropped" is more actionable when paired with "the producing service had a P1 incident 6 hours ago."

References

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions